Problem 1165. Convert double scalar to half-precision floating point (IEEE 754r)
Use MATLAB to convert a scalar double into a half-precision floating point. The return value should be a uint16.
The half-precision floating point format is specified here, and is the source for much of the test suite:
http://en.wikipedia.org/wiki/Half-precision_floating-point_format
This is an implementation in C if you want some inspiration to get started:
As other problems related to half-precision are added, I will try to link them here.
Solution Stats
Problem Comments
-
1 Comment
The only thing missing from the test suite are nans,regardless, it's a very good problem. Learned how to extract minus from minus 0, and finally took my time to learn about the controversial denormal floating-points numbers.
Solution Comments
Show commentsProblem Recent Solvers6
Suggested Problems
-
5852 Solvers
-
Find relatively common elements in matrix rows
2037 Solvers
-
2485 Solvers
-
Map all the indices of an Array Indices into a Vector giving Index vs Row and Column
41 Solvers
-
Combinations without using nchoosek
125 Solvers
More from this Author2
Problem Tags
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!